Meta is introducing facial recognition once more. Indeed, that very same technology that made privacy enthusiasts cringe back in 2021. Today, it's used to combat what are referred to as "celeb-bait" scams, where fake adverts use images of celebrities in an attempt to deceive. The twist,you might need to let Meta compare a face from an ad to profile pics of public figures. They say it serves to filter out impersonators. And yes, there are opt-in options. But does that make us safer, or more paranoid?
Testing in U.K., EU & South Korea
Meta began testing this stuff in some places to begin with. In the UK and EU, they recruited regulators. Public figures will shortly receive notifications inviting them to a "celeb-bait protection" tool. What this implies is those popular people are able to decide to allow Meta to scan ads that feature their face and determine if an ad is legitimate or fake. The same technology is being implemented to assist users in recovering hijacked accounts with video selfies. Optional, likely because requiring people to upload video selfies is a lawsuit waiting to happen.
Meta assures that facial information will be used only for comparisons and then discarded. That's an improvement on "we might hold onto it" but still makes eyebrows rise. Particularly since Meta recently closed down face recognition in 2021 due to privacy backlash and then settled cases for biometric abuses. Those memories are still recent.
Does It Work, Or Just Look Like It Works?
Meta says some numbers are trending how they want; user complaints of celebrity-bait ad fraud fell by 22%. Their AI is catching ever more before humans even get to see them. They say the facial recognition increase doubled or more than doubled the number of fake celeb ads that their tests picked up. And that now close to 500,000 public figures are being shielded from having their faces exploited. That's a big group of individuals.
Nevertheless, "fell by 22%" does not equal "disappeared." Scammers are clever. Phony celeb mugshots plus generative AI equals an arms race. Meta can filter out a lot, but new scams can always get through. Plus, even if it's "voluntary," there's always a pressure component. Public figures who decline may feel exposed. And users who are requested to send video selfies may fear what else Meta will do with the information. There’s a long history there.
Privacy, Laws & Trust Issues
Then again, in 2021, Meta shut down its facial recognition features after criticism and demands that it had broken biometric privacy regulations. Then there were suits. Then settlements. Then suspicion. And if that wasn't enough, regulators in the UK and EU are extremely wary of face ID technology. So Meta is attempting to do this carefully, with voluntary features, erasing face data after single uses, informing public figures, and cooperating with regulators. But trust isn't established overnight.
In other nations, laws require severe protection of data: GDPR in the EU, privacy regulations in the UK, and others in South Korea. So any slip could cost money or worse. Individuals are concerned about misuse, creep of surveillance, false positives, or choices by algorithmic guessing without human common sense. Meta claims it's conducting risk audits, complying with laws, and erasing data when finished. But words are simple; follow-through is difficult.
Progress With a Side of Skepticism
Meta’s celeb-bait face ID stuff is a mixed bag. On one side, it's good that they're trying something to stop scams that trick people. On the other side, we’re talking about face data, identity, and everything that comes with “optional” tools when big companies do them. The past is full of promises and missteps.
If this succeeds as promised, we could see fewer celeb-lookalike fake ads, more secure users, and improved account recovery tools. But we should also continue asking, "How much data is stored? Who looks at it? What about folks who say 'no' but are pressured nonetheless?" Because in tech, ethics always arrive after the party's started.
